How menu engineering works
Most café owners know their best sellers. Fewer know which of those best sellers actually make money. Menu engineering puts both questions side by side for every item:
- Profit per item: the price before tax minus the food cost. An item is high-profit if it earns more than the group's average, weighted by how many of each you sell.
- Popularity: the item's share of everything sold in the group. It counts as popular if its share is at least 70% of an equal share. With eight brunch dishes, that's 8.75%.
The method dates back to Michael Kasavana and Donald Smith's work in the 1980s and it's still the standard approach. The 70% rule stops a long menu from making nearly everything look unpopular.
The four categories
| High profit per item | Low profit per item | |
|---|---|---|
| Popular | Star Protect and promote | Plough horse Reprice or re-cost |
| Unpopular | Puzzle Reposition and sell | Dog Remove or rework |
A worked example: an all-day brunch menu
The calculator opens with eight brunch dishes from a four-week period. 1,060 were sold, the weighted average profit is $13.54 a dish, and food cost across the group is 28.9%.
- Stars: eggs benedict (240 sold, $15.20 profit each) and the big breakfast (150 sold, $15.40). Keep them exactly as they are and put them first on the menu board.
- Plough horses: smashed avocado (220 sold, $13.10) and the BLT (180 sold, $10.40). Both sell well but earn less than average. A dollar on the BLT is worth about $180 a month; a cheaper bread or a slightly smaller avocado portion does the same job without a price change.
- Puzzles: corn fritters ($15.60 profit) and the pancake stack ($14.80) earn more than most dishes but sell only 60 and 90. They're worth a better spot on the board, a photo on social media, or a staff recommendation. The pancakes miss the popularity line by a whisker.
- Dogs: mushrooms on toast and the granola bowl. Before cutting them, ask whether they serve a purpose. If the granola bowl is the one option for customers avoiding bread and eggs, keep it, cost it tightly and move it down the board.
Change the numbers to your own and the picture will be different. That's the point.
Getting reliable results in a café
- Keep coffee out of it. Coffee sells in such volume it makes every food item look unpopular. Run drinks as their own group if you want to engineer them.
- Separate brunch from cabinet food. A $6 slice will never earn what a $21 eggs benedict does. Mix them and every cabinet item looks like a Dog.
- Use four to eight weeks of sales. Your POS item report has the numbers. A shorter window gets thrown around by weather and school holidays.
- Count cabinet waste in the cost. A cabinet item that sells well but gets binned at close every second day costs more than its recipe says. The café menu price calculator builds that allowance in.
- Use prices excluding GST. The tax isn't profit.
Acting on the results
Change one thing at a time and re-run the numbers after the next four weeks. Move a Puzzle to the top of the board, or put a dollar on a Plough horse, then check what happened. A menu reprint is a good moment to do this properly, but most of the changes, like where an item sits on the board and what staff recommend, cost nothing.
For the strategy behind the prices, see how to price your café menu for profit. For bringing down food cost across the board, see how to reduce café food cost. For lifting what each customer spends, see how to increase average transaction value.
Frequently asked questions
What is menu engineering for a café?
Sorting every item on the menu by how well it sells and how much profit it makes per sale, then treating each group differently: promote Stars, reprice Plough horses, reposition Puzzles and remove or rework Dogs.
Should I run coffee through menu engineering?
Run it separately. Coffee sells in such volume that it would make every food item look unpopular. Drinks, brunch and cabinet food each work best as their own group.
How is the popularity threshold calculated?
Divide 100% by the number of items, then take 70% of that. With eight brunch dishes, any dish with at least 8.75% of brunch sales counts as popular.
Where do I get the numbers from?
Your POS item sales report for the last four to eight weeks gives the number sold and price — a system like TouchBistro pulls that straight off your till instead of you exporting and re-sorting it by hand. Food cost per item comes from your recipe costings. The café menu price calculator will cost an item in a couple of minutes if you don't have one. (Affiliate link: we may earn a commission if you book a demo through it, at no extra cost to you.)